Privacy policies like California Consumer Privacy Act (CCPA), App Tracking Transparency (ATT) framework from Apple and the General Data Protection Regulation (GDPR) enforced by the European Union (EU) are limiting third-party cookie advertising, resulting in more restricted audiences for advertisers to engage with. This has led eCommerce companies to shift emphasis to first-party data collection as a sustainable long-term strategy. The process usually generates a smaller and potentially less targeted customer database, leaving room for eCommerce platforms to design new marketplaces to trade the collected data (in)efficiently at a profit. In this paper, we propose a (first-party) data trading model where an (e-commerce) platform mediates interactions between many (finite) sellers and many (unit mass) heterogeneous buyers. We solve for a revenue-maximizing trading mechanism and study how initial endowments of first-party data affect its design. The optimal mechanism sells data in <em>bundles</em> to extract higher rents. By treating sellers symmetrically with regards to their (virtual) valuations, irrespective of the differences in initial endowments, makes the bundling strategy more profitable. Lastly, (ex-ante) sellers may be better-off withholding some of the data.